yutha_bible_tamil

LoRA adapter for Tamil-English Bible text generation, fine-tuned on Qwen3-8B.

Model Details

Model Description

This model is a LoRA adapter trained on Tamil-English Bible verses for biblical text generation and cross-lingual Bible study assistance.

  • Developed by: Sujith S
  • Model type: LoRA Adapter
  • Language(s): Tamil, English
  • License: Apache 2.0
  • Finetuned from model: unsloth/qwen3-8b-unsloth-bnb-4bit

Model Sources

Training Details

Training Data

Tamil-English Bible dataset containing all Bible verses.

Training Hyperparameters

  • Learning Rate: 2e-4
  • Batch Size: 2 per device (effective: 8)
  • Gradient Accumulation Steps: 4
  • Epochs: 3
  • Max Sequence Length: 2048
  • Optimizer: AdamW 8bit
  • Training regime: BF16/FP16 mixed precision

Speeds, Sizes, Times

  • Training Duration: 3 days
  • Adapter Size: 174MB

Intended Use

  • Tamil Bible text generation
  • Biblical content creation in Tamil
  • Cross-lingual Bible study assistance

Limitations

  • Specialized for biblical content
  • May not perform well on general Tamil text
  • Requires base model to function

Model Card Authors

Sujith S

Model Card Contact

rexmi.in

Framework versions

  • PEFT 0.16.0
Downloads last month
18
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support